A real agreement between real parties is powerful evidence of something. Establishing that it is evidence of this is the work.
Start a conversation with the IP Concierge, already scoped to comparable licenses. Pick a starting point, or describe the matter directly.
Actual licenses are the most persuasive input available to a damages analysis, because they record what parties genuinely agreed rather than what an economist reconstructs. That persuasiveness is exactly why courts examine them closely. A license offered as comparable has to be shown comparable — technically, in that it covers similar technology, and economically, in that it arose in circumstances resembling the hypothetical negotiation in scope, field, term and posture. The recurring failure is not selecting bad licenses; it is asserting comparability rather than establishing it, and then drawing from an agreement a proposition it does not support. Lump-sum agreements are where this bites hardest, because converting a single payment into a per-unit rate requires evidence about what the parties understood, and that evidence is often simply absent.
The dimensions a court examines, each of which can defeat reliance on its own.
Whether the licensed technology is genuinely similar to the asserted claims, assessed by someone competent to say so.
Scope, exclusivity, field of use, territory, term and the commercial relationship between the parties.
What else the agreement conveyed. A rate for forty patents says little about one of them without more.
A single payment does not convert into a per-unit rate; treating it as one requires support in the record.
Agreements struck to end litigation carry litigation pressure in their price, which affects what they can show.
Distance from the hypothetical negotiation date, and what had changed in the market between the two.
How comparability is established rather than asserted.
Usually the single largest input to the damages case.
Agreements a company signed years ago, on terms nobody thought would matter, become the ceiling on what it can credibly claim later. Licensing programs are worth running with the knowledge that every agreement is a future exhibit — including the recitals, and including what was bundled in to close the deal.
Sometimes, and it requires care in both directions. Agreements reached to end litigation reflect the cost and risk of litigation as well as the value of the technology, which courts recognize and which affects the weight they carry. They are not categorically inadmissible, and in some fields they are nearly all the evidence there is. What they cannot do is stand unexplained: an expert relying on one needs to address the litigation context rather than presenting the number as a market rate.
Then the analysis needs evidence that allows the asserted patent to be separated from the rest, and this is the failure mode that produced the most striking recent outcome — an appellate court in 2025 affirmed the exclusion of an expert who relied on a portfolio license without apportioning the value of the asserted patent from the others covered, and the resulting award was nominal. The separation does not have to be perfect, but it has to be reasoned from evidence rather than assumed or split evenly for convenience.
With difficulty, and only with support. Deriving an implied per-unit royalty from a single payment requires knowing what volume the parties had in mind, and unless the agreement or the negotiation record says so, that figure is being supplied by the expert rather than by the evidence. Recent authority has been pointed about this: testimony that both parties understood a lump sum to reflect a particular per-unit rate needs something in the record establishing that understanding.
It can, if the differences are addressed rather than ignored. Comparability is a matter of degree, and courts accept adjustment where an expert explains the basis for it. What fails is silent adjustment — applying a rate from a different technology, market or scope without acknowledging the gap, or applying an unexplained discount that appears calibrated to reach a desired result.
Describe the licensing history. The Institute will help you see what those agreements can and cannot establish.